Competition Law and Future Institutional Evolution of Antitrust in Intelligent Civilizations

Introduction

The expression “intelligent civilizations” describes an economic environment in which artificial intelligence, autonomous agents, algorithmic decision-making, robotics, connected platforms, cloud infrastructure, advanced data systems and machine-to-machine commerce increasingly participate in economic activity.

In such an environment, competition law cannot remain confined to the traditional model of examining human firms competing in identifiable product markets. Market power may increasingly arise from control over computational infrastructure, foundation models, data, chips, cloud capacity, algorithms, interfaces, digital identity, autonomous-agent ecosystems and standards.

The institutional evolution of antitrust therefore concerns not merely changing legal rules, but changing the capacity, structure and methods of competition authorities themselves.

Recent regulatory work already reflects this transition. For example, the UK CMA's work on AI foundation models has identified risks involving control of critical inputs, incumbent leverage across markets and partnerships that may reinforce market power. International competition authorities have also expressly recognized the need for cooperation concerning generative-AI foundation models and AI products.

I. Meaning of Institutional Evolution of Antitrust

Institutional evolution means the gradual transformation of competition authorities from conventional enforcement agencies into technology-intensive, continuously monitoring, cross-sector and internationally coordinated institutions.

Traditional antitrust institutions generally perform:

1. merger review;
2. cartel investigation;
3. abuse-of-dominance investigations;
4. market studies;
5. economic analysis;
6. remedies and enforcement.

Intelligent markets add new institutional requirements:

• algorithmic auditing;
• AI-system expertise;
• computational economics;
• data governance;
• real-time market monitoring;
• technical interoperability analysis;
• cloud and compute expertise;
• cybersecurity expertise;
• digital-forensics capabilities;
• cross-border cooperation;
• continuous assessment of emerging technologies.

Thus, the future competition authority may resemble a combination of:

antitrust agency + technology regulator + economic observatory + data laboratory + digital-forensics institution.

II. Why Intelligent Markets Challenge Traditional Antitrust Institutions

1. Markets become multidimensional

A conventional market may be analysed through:

Product → Geography → Competitors → Prices → Market share

An intelligent market may require:

Data → Compute → Models → APIs → Platforms → Agents → Users → Ecosystem → Complementary services

A company may have relatively modest market share in one layer while exercising substantial strategic control over another.

For example:

• a cloud provider may control computing capacity;
• an AI developer may control a foundation model;
• a platform may control distribution;
• an operating system may control access to users;
• an app store may control monetisation;
• a payment network may control transactions.

The institutional problem is therefore vertical and ecosystem power, rather than simply horizontal market share.

III. Evolution from Ex Post to Ex Ante Antitrust

Traditional antitrust frequently operates after harmful conduct has occurred.

Intelligent markets create circumstances where waiting for harm may be costly because:

• network effects can become irreversible;
• data advantages can compound;
• AI models improve through scale;
• switching costs can increase;
• ecosystems can rapidly acquire users;
• acquisitions can eliminate emerging competitors before they become significant.

Consequently, future authorities may need a combination of:

Ex post enforcement

Investigate conduct after suspected infringement.

Ex ante supervision

Establish obligations before harmful conduct becomes entrenched.

Continuous monitoring

Observe rapidly changing markets rather than relying exclusively on periodic investigations.

This represents a major institutional transformation.

IV. From Market Definition to Ecosystem Mapping

Traditional market definition remains important, but intelligent ecosystems require broader mapping.

A competition authority may need to identify:

Layer 1 — Physical infrastructure

• semiconductor manufacturing;
• data centres;
• energy;
• telecommunications.

Layer 2 — Computational infrastructure

• GPUs;
• specialised AI accelerators;
• cloud computing;
• distributed computing.

Layer 3 — Data

• training data;
• behavioural data;
• proprietary datasets;
• real-time data.

Layer 4 — Foundation models

• large language models;
• multimodal models;
• specialised models.

Layer 5 — Platforms

• search;
• social networks;
• marketplaces;
• operating systems.

Layer 6 — AI agents

• autonomous purchasing;
• financial agents;
• logistics agents;
• software-development agents.

Layer 7 — End-user markets

The authority must understand how power propagates between these layers.

V. Future Institutional Model

A future competition authority could develop specialised divisions.

1. AI Competition Division

Its responsibilities could include:

• foundation-model markets;
• algorithmic exclusion;
• AI partnerships;
• model-access restrictions;
• AI-enabled collusion;
• algorithmic discrimination;
• autonomous-agent competition.

The CMA has already developed dedicated expertise around AI foundation models and identified competition risks across the AI value chain.

2. Algorithmic Audit Division

Competition authorities may increasingly need to inspect:

• pricing algorithms;
• ranking algorithms;
• recommendation systems;
• bidding algorithms;
• search algorithms;
• allocation algorithms.

The purpose would not be to regulate algorithms merely because they are sophisticated, but to determine whether they facilitate:

• collusion;
• exclusion;
• discrimination;
• self-preferencing;
• foreclosure;
• manipulation.

3. Data Competition Division

Data may function as an essential competitive input.

The authority could examine:

• exclusive data arrangements;
• data portability;
• data interoperability;
• discriminatory data access;
• data pooling;
• data-driven exclusion;
• acquisition of data-rich companies.

4. Computational Infrastructure Division

Future antitrust institutions may need specialists capable of analysing:

• GPU concentration;
• cloud infrastructure;
• semiconductor supply;
• model-training capacity;
• computing bottlenecks;
• cloud switching costs.

This becomes particularly important because AI competition may depend not merely on intellectual property but on access to sufficient computing resources.

VI. Six Major Case Laws Demonstrating the Institutional Evolution

1. United States v. Microsoft Corp. (2001)

Background

Microsoft was accused of using its dominance in PC operating systems to restrict competition in internet browsers.

Competition principle

The case demonstrated the danger of leveraging power in one technological layer into another.

Institutional significance

The case showed that competition authorities must understand:

• technological architecture;
• operating-system design;
• software interfaces;
• network effects;
• platform leverage.

Future relevance

AI ecosystems may reproduce the same structure:

Operating system → cloud → AI assistant → applications → users

The institutional lesson is that authorities must analyse technological ecosystems rather than isolated products.

2. United States v. Apple / Epic Games v. Apple

The Apple litigation surrounding App Store restrictions illustrates another institutional problem: control over an important digital distribution infrastructure.

Issues included:

• platform governance;
• app distribution;
• payment systems;
• commissions;
• restrictions on alternative payment mechanisms;
• platform rules.

Institutional significance

Competition authorities increasingly need expertise in platform governance, not merely pricing.

In intelligent economies, an AI platform may similarly determine:

• which models can be distributed;
• which agents receive access;
• what APIs are available;
• what payments are permitted;
• which applications receive visibility.

Thus, future antitrust institutions must understand technical gatekeeping.

3. Google Shopping — European Commission / General Court

The Google Shopping litigation concerned Google's treatment of comparison-shopping services within its search ecosystem.

Competition principle

The case illustrates how dominance in one digital service can potentially be leveraged to favour another service operated by the same undertaking.

Institutional importance

Digital competition requires investigation of:

• search rankings;
• algorithms;
• visibility;
• traffic allocation;
• platform neutrality;
• self-preferencing.

Future relevance

AI assistants could become the principal gateway through which consumers obtain information.

An AI agent that controls recommendations could potentially influence:

which seller → which product → which service → which payment provider

gets selected.

The future competition authority therefore needs capacity to audit machine-mediated choice architecture.

4. Google Android — European Commission

The Android case involved Google's conduct concerning the Android ecosystem, including contractual arrangements relating to search, browsers and app distribution.

Competition principle

The case demonstrates the significance of:

• tying;
• default arrangements;
• ecosystem leverage;
• contractual restrictions;
• network effects.

Institutional significance

The relevant competitive unit was not simply an individual product. The investigation required analysis of the entire mobile ecosystem.

Future application

AI ecosystems could involve:

Cloud + model + operating system + assistant + app marketplace + payments.

A future competition authority therefore needs ecosystem-wide investigative capabilities.

5. Qualcomm Antitrust Litigation

Qualcomm-related antitrust litigation and enforcement across jurisdictions illustrates the competition problems surrounding technology licensing and essential technological inputs.

Important issues include:

• patents;
• licensing;
• royalties;
• chipsets;
• vertical integration;
• exclusionary strategies;
• technological standards.

Institutional significance

Technology markets frequently combine:

IP + standards + hardware + software + licensing.

Future AI competition may similarly involve:

• model licences;
• AI standards;
• accelerator technology;
• proprietary APIs;
• interoperability rights.

Authorities therefore need personnel who understand both competition law and technological intellectual property systems.

6. FTC v. Amazon

The FTC and state attorneys general sued Amazon in 2023, alleging that Amazon used interconnected strategies to maintain monopoly power and restrict competition. The allegations include practices affecting sellers, prices, product visibility and rival competition.

Institutional significance

The case demonstrates the increasing importance of analysing multiple interconnected practices rather than isolated conduct.

This is especially important in intelligent economies because an AI-enabled platform may simultaneously control:

• marketplace access;
• advertising;
• logistics;
• seller information;
• recommendation systems;
• pricing tools.

Future lesson

Competition authorities need the ability to examine cumulative ecosystem effects.

VII. Additional Important Case Laws

7. United States v. Google — Search and Advertising

The Google litigation concerning search and digital advertising demonstrates the increasing importance of:

• defaults;
• distribution agreements;
• advertising technology;
• data;
• search infrastructure;
• digital ecosystems.

It highlights the institutional need for authorities capable of analysing markets in which data, attention and distribution are economically significant.

8. Ohio v. American Express Co.

The Supreme Court's decision concerned a two-sided transaction platform and the analysis of effects across both sides of that platform.

Institutional importance

Digital markets often operate through:

Users ↔ Platform ↔ Merchants

or:

Advertisers ↔ Platform ↔ Consumers

The case illustrates why future competition authorities require sophisticated multi-sided-market economics.

VIII. From Human Coordination to Machine Coordination

One of the most important future challenges is algorithmic coordination.

Suppose competing AI systems independently observe:

• prices;
• demand;
• inventory;
• competitors' behaviour.

They may adjust prices automatically.

Traditional cartel law generally looks for human communication or coordinated conduct. Intelligent markets raise a harder question:

What happens when coordination emerges from autonomous systems without conventional human communication?

This requires institutions capable of distinguishing:

1. legitimate independent algorithmic optimisation;
2. conscious coordination;
3. hub-and-spoke coordination;
4. algorithm-assisted collusion;
5. tacit coordination;
6. autonomous strategic interaction.

The CMA has previously examined how algorithms can affect competition and consumer outcomes, demonstrating why technical expertise is increasingly necessary.

IX. AI Partnerships as an Institutional Challenge

AI development frequently involves partnerships between:

• cloud companies;
• semiconductor companies;
• model developers;
• platforms;
• investors;
• application providers.

A partnership may not look like a traditional merger.

Nevertheless, it can potentially influence:

• access to compute;
• model distribution;
• data;
• investment;
• technical standards;
• downstream markets.

The CMA has specifically identified partnerships involving key players as a potential mechanism through which existing positions of market power could be reinforced across the AI value chain.

This suggests that future competition authorities may require a transactional ecosystem review system covering not only mergers but also strategically important partnerships.

X. Continuous Market Monitoring

Traditional competition investigations can take years.

Intelligent markets may change within months or even weeks.

Future authorities may therefore maintain:

Digital market observatories

They could continuously monitor:

• market shares;
• prices;
• API access;
• switching rates;
• acquisitions;
• model availability;
• compute concentration;
• interoperability;
• algorithmic behaviour.

The objective would be early detection, rather than waiting until market foreclosure becomes difficult to reverse.

XI. Competition Authority as a Data Institution

Future authorities may need access to:

• anonymised transaction datasets;
• platform logs;
• algorithmic records;
• API records;
• pricing data;
• model-performance information;
• merger databases.

They could establish secure regulatory data laboratories.

Regulatory technology could include:

Data ingestion → anomaly detection → economic modelling → algorithmic testing → human investigation → enforcement

This represents a transformation from a predominantly legal institution into a legal-economic-computational institution.

XII. International Institutional Evolution

Intelligent markets are inherently transnational.

A single AI system may involve:

• training in one country;
• cloud infrastructure in another;
• users worldwide;
• data from multiple jurisdictions;
• corporate headquarters elsewhere.

Consequently, national competition authorities increasingly require:

• information-sharing arrangements;
• coordinated investigations;
• compatible merger-review standards;
• cross-border dawn-raid cooperation;
• technical knowledge-sharing.

In July 2024, competition authorities from the US, EU and UK issued a joint statement concerning competition in generative-AI foundation models and AI products, demonstrating this emerging cooperative institutional model.

XIII. Future Institutional Architecture

A possible institutional architecture can be represented as follows:

                    FUTURE ANTITRUST AUTHORITY                              │              ┌───────────────┼────────────────┐              │               │                │        Legal Division   Economics Unit   Technology Unit              │               │                │              └───────────────┼────────────────┘                              │                     Digital Markets Lab                              │          ┌───────────────────┼───────────────────┐          │                   │                   │      AI/Algorithms        Data/Cloud          Cyber/Forensics          │                   │                   │          └───────────────────┼───────────────────┘                              │                     Market Observatory                              │                    Cross-Border Cooperation                              │                       Enforcement

XIV. From Competition Authority to Competition Ecosystem Authority

The future institution may have five interconnected functions.

1. Detection

Identify emerging risks.

2. Diagnosis

Determine whether conduct creates competitive harm.

3. Prediction

Assess whether current conduct could produce durable foreclosure.

4. Intervention

Use behavioural or structural remedies.

5. Monitoring

Determine whether remedies continue to work.

This produces a continuous regulatory cycle:

Observe → Analyse → Investigate → Intervene → Monitor → Reassess

XV. New Institutional Powers That May Become Necessary

Future authorities may require carefully defined powers relating to:

A. Algorithmic access

Authority to inspect relevant algorithmic systems where legally justified.

B. Data access

Controlled access to relevant datasets.

C. Technical testing

Ability to conduct independent experiments on digital platforms.

D. Interoperability

Assessment of technical barriers preventing competitors from connecting.

E. Rapid interim measures

Temporary intervention where network effects could make eventual remedies ineffective.

F. Merger and acquisition monitoring

Closer scrutiny of acquisitions of emerging AI firms and strategically important technologies.

G. Partnership review

Examination of non-merger arrangements that may materially alter competitive structure.

XVI. Risks of Excessive Institutional Expansion

Institutional evolution must also respect limits.

Competition authorities should not become general-purpose technology regulators without clear statutory authority.

Potential problems include:

• regulatory duplication;
• excessive compliance costs;
• reduced innovation;
• conflicting regulatory objectives;
• politicisation;
• insufficient procedural safeguards;
• excessive disclosure of trade secrets;
• uncertainty for businesses.

Therefore, future institutional design should maintain:

independence + transparency + expertise + due process + judicial review.

XVII. Remedies in Intelligent Economies

Traditional remedies may increasingly be supplemented by technological remedies.

Behavioural remedies

• non-discrimination;
• transparent ranking;
• access obligations;
• data portability;
• interoperability.

Structural remedies

• divestiture;
• separation of business units;
• ownership restrictions.

Technical remedies

• API access;
• interoperability;
• portability;
• technical separation;
• neutral interfaces.

Data remedies

• controlled data access;
• data portability;
• data-sharing obligations where legally appropriate.

The remedy must correspond to the mechanism through which competitive harm occurs.

XVIII. Future Role of Competition Economists

Competition economics itself will evolve.

Traditional indicators such as:

• price;
• output;
• market share;
• margins;

will remain important.

But authorities may increasingly analyse:

• compute concentration;
• data concentration;
• model quality;
• switching costs;
• user attention;
• API dependency;
• model interoperability;
• innovation rates;
• ecosystem centrality;
• agent adoption.

The central question may move from:

“What is the firm's market share?”

to:

“What economic bottlenecks does the firm control, and how can that control affect competitive entry and innovation across the ecosystem?”

XIX. Future Competition Authorities and Intelligent Agents

A particularly important development will be the emergence of AI purchasing agents.

Consumers may increasingly delegate decisions to autonomous agents.

Instead of:

Consumer → Search → Compare → Purchase

the process may become:

Consumer → AI Agent → Multiple Platforms → Selection → Purchase

This could change competition profoundly.

If one AI agent becomes the dominant intermediary, it could influence:

• product visibility;
• prices;
• supplier selection;
• advertising;
• payment systems;
• logistics.

Therefore, future antitrust institutions may need to examine agent-mediated markets separately from conventional consumer markets.

XX. Core Legal Doctrines Likely to Evolve

The following doctrines may require reinterpretation or supplementation:

XXI. Six Core Lessons from the Case Law

The major cases collectively demonstrate six institutional lessons:

1. Microsoft

Competition authorities must understand technology architecture.

2. Google Shopping

Authorities must understand algorithmic visibility and ranking.

3. Google Android

Authorities must understand ecosystem leverage and defaults.

4. Qualcomm

Authorities must understand technology standards, licensing and infrastructure.

5. Amazon

Authorities must understand interconnected platform strategies.

6. American Express

Authorities must understand multi-sided markets and indirect competitive effects.

Together, these cases demonstrate why future antitrust institutions cannot rely exclusively upon traditional legal and economic tools.

XXII. Proposed Model for Institutional Evolution

A mature intelligent-economy competition authority could operate through seven institutional pillars:

Pillar 1 — Legal capacity

Competition lawyers specialising in digital markets, IP, contracts and technology.

Pillar 2 — Economic capacity

Economists capable of modelling dynamic and multi-sided markets.

Pillar 3 — Technical capacity

AI engineers, data scientists, cybersecurity specialists and algorithm auditors.

Pillar 4 — Market intelligence

Permanent monitoring of emerging technologies and market structures.

Pillar 5 — International cooperation

Real-time cooperation with foreign competition authorities.

Pillar 6 — Procedural safeguards

Due process, transparency, confidentiality and judicial review.

Pillar 7 — Adaptive remedies

Remedies capable of changing as technology changes.

Conclusion

The future institutional evolution of antitrust in intelligent civilizations is ultimately a transition from static enforcement toward adaptive competition governance.

The traditional competition authority asks:

Who has market power, what conduct occurred, and what harm resulted?

The intelligent-economy competition authority will increasingly need to ask:

Who controls the critical technological infrastructure, how does that control propagate across interconnected ecosystems, and what institutional intervention is necessary to preserve contestability and innovation?

The evolution is therefore likely to move:

Firm-centric → ecosystem-centric

Human decision-making → human-machine decision-making

Periodic investigation → continuous monitoring

Price analysis → data, compute, quality and innovation analysis

Ex post enforcement → combined ex ante and ex post supervision

National enforcement → international institutional cooperation

Legal expertise → legal + economic + computational expertise

The ultimate institutional objective is not to prevent technological concentration as such. It is to ensure that technological advancement does not make markets structurally uncontestable, while preserving incentives for investment, innovation and legitimate technological integration.